Assessment of accuracy of laboratory testing results, relative to peer group consensus values in external quality control, by bivariate z-score analysis: the example of D-Dimer.

OBJECTIVES: The aim of this study is to develop a practical method for bivariate z-score analysis which can be applied to the survey of an external quality assessment programme. METHODS: To develop the bivariate z-score analysis, the results of four surveys of the international D-Dimer external quality assessment programme of 2022 of the ECAT Foundation were used. The proposed methodology starts by identifying the bivariate outliers, using a Supervised Sequential Hotelling T2 control chart. The outlying data are removed, and all the remaining data are used to provide robust estimates of the parameters of the assumed underlying bivariate normal distribution. Based on these estimates two nested homocentric ellipses are drawn, corresponding to confidence levels of 95 and 99.7 %. The bivariate z-score plot described provides the laboratory with an indication of both systematic and random deviations from zero z-score values. The bivariate z-score analysis was examined within survey 2022-D4 across the three most frequently used methods. RESULTS: The number of z-score pairs included varied between 830 and 857 and the number of bivariate outliers varied between 20 and 28. The correlation between the z-score pairs varied between 0.431 and 0.647. The correlation between the z-score pairs for the three most frequently used varied between 0.208 and 0.636. CONCLUSIONS: The use of the bivariate z-score analysis is of major importance when multiple samples are distributed around in the same survey and dependency of the results is likely. Important lessons can be drawn from the shape of the ellipse with respect to random and systematic deviations, while individual laboratories have been informed about their position in the state-of-the-art distribution and whether they have to deal with systematic and/or random deviations.

[Long-term automated Bayesian management of IQC results: The experience of the Lyon Hospitals Board hemostasis laboratory]. / Déploiement d'une gestion automatisée au long cours des résultats de CIQ en approche bayésienne : retours d'expérience du Laboratoire d'hémostase aux Hospices Civils de Lyon.

The Lyon Hospitals Board (HCL) hemostasis laboratory has shifted from a frequentist to a long-term Bayesian approach to IQC results management, using the Hemohub® software of the Werfen corporation, which hosts the requisite Bayesian tools. IQC plans based on supplier specifications proved effective in managing analytic risk in line with the ISO 15189 standard. Long-term Hemohub® control and monitoring has been validated by acceptable feedback from the EQA organization used by the hemostasis community.

Cost-effectiveness of thrombotic thrombocytopenic purpura diagnosis: a retrospective analysis in the University Hospital Center of Lyon (France).

The aim of the present study was to perform an economic evaluation of two alternative assays of ADAMTS13 activity (A Disintegrin And Metalloprotease with ThromboSpondin type 1 repeats, member 13) for diagnosing thrombotic thrombocytopenic purpura (TTP) in the Hospital of Lyon (France). The study approach was more economic than clinical. We retrospectively calculated the prescription costs of ADAMST13 activity from January to December 2019 for patients depending on the assay: manual ELISA (Technozym) or automated assay (AcuStar Werfen, Instrumentation Laboratory). Then, we compared the cost of therapeutic plasma exchange (TPE) consumption awaiting ADAMTS13 activity assay results. From an economic point of view, the automated assay was more cost-effective. From a clinical one, we supposed that the faster results given by AcuStar could improve patient care by reducing the number of TPEs. Automated assay could improve patient care without increasing costs in our institution.

Prospective evaluation of two specific IgG immunoassays (HemosIL® AcuStar HIT-IgG and HAT45G® ) for the diagnosis of heparin-induced thrombocytopenia: A Bayesian approach.

INTRODUCTION: The accurate diagnosis of heparin-induced thrombocytopenia (HIT) is essential to ensure adequate treatment and prevent complications. First step diagnosis test are immunoassays including enzyme-linked immunosorbent assays (ELISAs) and rapid immunoassays. METHODS: Using a Bayesian approach, we prospectively evaluated the performance of the IgG PF4/polyvinylsulfonate ELISA and a chemiluminescent immunoassay (CLIA), which are specific for IgG and use the same antigenic target to detect HIT antibodies. RESULTS: One hundred and eighty-four 184 consecutive patients with an intermediate (n = 159) or high (n = 25) clinical pretest probability of HIT based on the 4Ts score or platelet pattern were included. Both immunoassays (IAs) were performed on all 184 samples, and definite HIT was confirmed with a positive serotonin release assay in 29 patients (12.7%). The sensitivity (Ss) and negative predictive value (NPV) of ELISA were excellent (100%) allowing HIT to be excluded with good confidence when the test was negative. In addition, the Ss and NPV of the CLIA equalled 93.1% and 98.6%, respectively, as it was negative in two definite HIT. When the CLIA was negative, the post-test probability of HIT was 0.7% in case of intermediate risk. Although there was excellent agreement between CLIA and ELISA results, the quantitative values provided by the two IAs were not correlated. CONCLUSION: AcuStar HIT® detects more than 90% of HIT, as do all rapid IAs, and appears to be a good tool for excluding HIT when the pretest probability is intermediate. A chemiluminescent signal higher than 10 IU/mL is highly predictive of definite HIT with a PPV of 100%.

[Theoretic and practical contribution of bayesian inference to statistical process control: The experience of the Lyon Hospitals Board hemostasis laboratory]. / Intérêts théoriques et pratiques de l'inférence bayésienne dans la maîtrise statistique des procédés : retours d'expérience du laboratoire d'hémostase aux Hospices civils de Lyon.

Laboratories need to set up effective overall management of their internal quality control (IQC) and external quality assessment (EQA) results as key elements in statistical process control. Quality targets need to be defined, with methods to ensure durable control with respect to the relevant specifications. The hemostasis laboratory of the Lyon Hospitals Board (HCL, Lyon, France) uses model 3 from the Milan consensus conference, which is the state of the art in terms of quality targets, and uses a common EQA provider supplying as many real patient samples as possible. Giving priority to adopted methods, the lab optimizes the use of manufacturers' prior data: maximum acceptable inter assay coefficient of variation (CV) and prior IQC target values. Bayesian inference brings the method under control with respect to the manufacturers' prior data without the need for a preliminary phase. It links the IQC and EQA plans by the maximum acceptable CVs defined by the manufacturer.

Factor VIII and IX assays for post-infusion monitoring in hemophilia patients: Guidelines from the French BIMHO group (GFHT).

Replacement therapy with plasma-derived or recombinant FVIII and FIX (pdFVIII/pdFIX or rFVIII/rFIX) concentrates is the standard of treatment in patients with haemophilia A and B, respectively. Measurement of factor VIII (FVIII:C) or factor IX (FIX:C) levels can be done by one-stage clotting assay (OSA) or chromogenic substrate assay (CSA). The French study group on the Biology of Hemorrhagic Diseases (a collaborative group of the GFHT and MHEMO network) presents a literature review and proposals for the monitoring of FVIII:C and FIX:C levels in treated haemophilia A and B patients, respectively. The use of CSA is recommended for the monitoring of patients treated with pdFVIII or rFVIII including extended half-life (EHL) rFVIII. Except for rFVIII-Fc, great caution is required when measuring FVIII:C levels by OSA in patients substituted by EHL-rFVIII. The OSA is recommended for the monitoring of patients treated with pdFIX or rFIX. Large discordances in the FIX:C levels measured for extended half-life rFIX (EHL-rFIX), depending on the method and reagents used, must lead to great attention when OSA is used for measuring FIX:C levels in patients substituted by EHL-rFIX. Data of most of recent studies, obtained with spiked plasmas, deserve to be confirmed in plasma samples of treated patients.

Effects of hemolysis, bilirubin, and lipemia interference on coagulation tests detected by two analytical systems.

INTRODUCTION: Interference on biological assays due to hemolysis, icterus, or lipemia (HIL) could represent a significant source of analytical errors leading to inaccurate interpretation of results. The aim of this study was to assess the HIL interference on prothrombin time (PT), activated partial thromboplastin time (aPTT), and fibrinogen, using mechanical and optical detection methods. METHODS: Control plasmas and plasmas from patients treated with vitamin K antagonists or unfractionated heparin, with or without HIL, were performed on two analytical detection systems in order to identify potential analytical biases. Whether HIL lead to significant biological interferences was also evaluated, and a cutoff point for HIL-induced analytical bias was determined. RESULTS: Hemolysis influenced PT and aPTT when hemoglobin was at 5 and 1.5 g/L in plasma, respectively. At 1.8 g/L, a positive relationship was found between the bias and the hemoglobin supernatant level only for fibrinogen measurement, using optical detection. For icteric interference, no significant bias was observed until a bilirubin concentration of 30 mg/dL. Lipamia (>500 mg/dL) led to analytical interference when using the optical analyzer. CONCLUSION: The present study detected analytical interferences such as lipemia (>500 mg/dL) on coagulation tests on the optical analyzer. We also found a biological impact on the results in case of hemolyzed sample: Fibrinogen was decreased when the hemoglobin level was superior to 1.8 g/L, PT was prolonged beyond 5 g/L, and aPTT was shortened beyond 1.5 g/L hemoglobin concentration, especially in patients treated with heparin. Above these thresholds, it is important not to give results that could influence the clinical decision.

[Factor IX assays in treated hemophilia B patients]. / Dosage de l'activité du facteur IX chez les patients hémophiles B substitués.

Replacement therapy with plasma-derived or recombinant FIX (pdFIX or rFIX) concentrates is the standard of treatment in patients with hemophilia B. The method predominantly used for measuring factor IX (FIX:C) levels is the one-stage clotting assay (OSA) but this method depends on the activated partial thromboplastin time (APTT) reagent and the coagulation analyzer used, and wide variations in the measurements of FIX recovery have been reported with some factor concentrates. The French study group on the biology of hemorrhagic diseases (a collaborative group of the GFHT and MHEMO network), presents a review of the literature and proposals for the monitoring of FIX:C levels in treated hemophilia B patients. The use of OSA calibrated with a plasma reference tested against the current FIX WHO International Standard is recommended for the monitoring of patients treated with pdFIX or rFIX. Chromogenic substrate assays (CSA) are adequate for the monitoring of patients treated with Rixubis®, but data available for Benefix® are currently too limited. For extended half-life rFIX (EHL-rFIX), large discordances in the FIX:C levels measured were evidenced, depending on the method and reagents used. Great attention is therefore required for measuring FIX:C levels by OSA in patients substituted by EHL-rFIX. Commercial kits for CSA are not equivalent, and although potentially useful, they are not validated for all EHL-rFIX. Most of recent studies reported data obtained with spiked plasmas, which deserve to be confirmed on plasma samples collected in treated patients.

[Factor VIII assays in treated hemophilia A patients]. / Dosage de l'activité du facteur VIII chez les patients hémophiles A substitués.

Replacement therapy with plasma-derived or recombinant FVIII (pdFVIII or rFVIII) concentrates is the standard of treatment in patients with hemophilia A. The reference method used for measuring factor VIII (FVIII:C) levels in patients treated by FVIII concentrates is the chromogenic substrate assay (CSA). However, the one-stage clotting assay (OSA) is predominantly used in current clinical practice, but this method depends on the activated partial thromboplastin time (APTT) reagent and the coagulation analyzer used, and wide variations in the measurements of FVIII recovery have been reported with some factor concentrates. The French study group on the biology of hemorrhagic diseases (a collaborative group of the GFHT and MHEMO network) presents a review of the literature and proposals for the monitoring of FVIII:C levels in treated hemophilia A patients. The use of CSA calibrated with a plasma reference tested against the current FVIII WHO (World Health Organization) International Standard is recommended for the monitoring of patients treated with pdFVIII or rFVIII including extended half-life (EHL) rFVIII. OSA are adequate for the monitoring of patients treated with pdFVIII or with most of rFVIII concentrates. However, preliminary comparison with CSA is mandatory before measuring FVIII:C by OSA in patients treated by Refacto AF®. For rFVIII-EHL, OSA are only acceptable for Elocta®. Great caution is therefore required when measuring FVIII:C levels by OSA in patients substituted by other EHL-rFVIII. Indeed, most of recent studies reported data obtained with spiked plasmas, which deserve to be confirmed on plasma samples collected in treated patients.

Pre-analytical effects of pneumatic tube system transport on routine haematology and coagulation tests, global coagulation assays and platelet function assays.

INTRODUCTION: Pneumatic tube system (PTS) in hospitals is commonly used for the transport of blood samples to clinical laboratories, as it is rapid and cost-effective. The aim was to compare the effects on haematology samples of a newly acquired ~2km-long PTS that links 2 hospitals with usual transport (non-pneumatic tube system, NPTS). METHODS: Complete blood cell count, routine coagulation assays, platelet function tests (PFT) with light-transmission aggregometry and global coagulation assays including ROTEM® and thrombin generation assay (TGA) were performed on blood samples from 30 healthy volunteers and 9 healthy volunteers who agreed to take aspirin prior to blood sampling. RESULTS: The turnaround time was reduced by 31% (p<0.001) with the use of PTS. No statistically significant difference was observed for most routine haematology assays including PFT, and ROTEM® analysis. A statistically significant, but not clinically relevant, shortening of the APTT after sample transport by PTS was found (mean±SD: 30s±1.8 vs. 29.5s±2.1 for NPTS). D-dimer levels were 7.4% higher after transport through PTS but were not discordant. A statistically significant increase of thrombin generation was found in both platelet poor- and platelet rich- plasma samples after PTS transport compared to NPTS transport. CONCLUSION: PTS is suitable for the transport of samples prior to routine haematology assays including PFT, but should not be used for samples intended for thrombin generation measurement.

Use of prior manufacturer specifications with Bayesian logic eludes preliminary phase issues in quality control: an example in a hemostasis laboratory.

The present study seeks to demonstrate the feasibility of avoiding the preliminary phase, which is mandatory in all conventional approaches for internal quality control (IQC) management. Apart from savings on the resources consumed by the preliminary phase, the alternative approach described here is able to detect any analytic problems during the startup and provide a foundation for subsequent conventional assessment. A new dynamically updated predictive control chart (PCC) is used. Being Bayesian in concept, it utilizes available prior information. The manufacturer's prior quality control target value, the manufacturer's maximum acceptable interassay coefficient of variation value and the interassay standard deviation value defined during method validation in each laboratory, allow online IQC management. An Excel template, downloadable from journal website, allows easy implementation of this alternative approach in any laboratory. In the practical case of prothrombin percentage measurement, PCC gave no false alarms with respect to the 1ks rule (with same 5% false-alarm probability on a single control sample) during an overlap phase between two IQC batches. Moreover, PCCs were as effective as the 1ks rule in detecting increases in both random and systematic error after the minimal preliminary phase required by medical biology guidelines. PCCs can improve efficiency in medical biology laboratories.

A comparison of the 12s rule and Bayesian approach for quality control: application to one-stage clotting factor VIII assay.

An ideal medical biology internal quality control (IQC) plan should both monitor the laboratory methods efficiently and implement the relevant clinical-biological specifications. However, many laboratories continue to use the 12s quality control rule without considering the high risk of false rejection and without considering the relationship of analytical performance to quality requirements. Alternatively, one can move to the Bayesian arena, enabling probabilistic quantification of the information coming in, on a daily basis from the laboratory's IQC tests, and taking into account the laboratory's medical and economic contexts. Using the example of one-stage clotting factor VIII assay, the present study compares frequentist (12s quality control rule) and Bayesian IQC management with respect to prescriber requirements, process start-up phase issues, and abnormal scenarios in IQC results. To achieve comparable confidence, the traditional 12s quality control rule requires more data than the Bayesian approach in order to detect an increase in the random or systematic error of the method. Moreover, the Bayesian IQC management approach explicitly implements respect of prescriber requirements in terms of calculating the probability that the variable in question lies in a given predefined interval: for example, the factor VIII concentration required after knee surgery in a hemophilia patient.

Analytic variability due to change of deficient plasma vials: application to one-stage clotting factor VIII assay.

The statistical process control required under International Organization for Standardization 15189 as well as economic considerations necessitates having robust methods that do not need systematic recalibration for each series of analyses. Using the concrete example of one-stage clotting factor VIII assay, we assessed the analytic variability specifically linked to changing factor VIII deficient plasma vials. The study used freeze-dried (Instrumentation Laboratory, Siemens, Stago and T-Coag) and frozen (Affinity Biologicals and Precision Biologic) factor VIII deficient plasmas. On the most widely recognized acceptability criteria and methods (i.e. those of Kallner et al. and Kasper et al.), the Stago and Instrumentation Laboratory plasmas require systematic recalibration at each vial changeover.

Impact of quality control matrix effect: application to the calculation of uncertainty of measurement in one-stage clotting factor VIII assay.

In Europe, the ISO 15 189 standard requires uncertainty of measurement to be calculated for all measurands. We calculated the analytical imprecision and bias of our factor VIII coagulometric assay method between 5 and 80 U/dl, using plasmas expected to be at 5, 30 and 80 U/dl of factor VIII. We implemented Meijer et al.'s [Clin Chem 2002; 48:1011-1015] long-term coefficient of variance, bias and also uncertainty of measurement calculations. Assessments used reference plasma diluted in severe haemophilic plasma, in immunodepleted factor VIII-deficient plasma and in bovine serum albumin. With plasmas diluted in severe haemophilic and immunodepleted factor VIII-deficient plasma, calculated uncertainty of measurement was 10% compared with 15% (i.e., 50% greater) for plasma diluted in albumin buffer or as calculated from European Concerted Action on Thrombosis consensus values. It is thus important to approximate the patient sample matrix to obtain as precise an estimation as possible of assay method uncertainty of measurement.

Bayesian logic in statistical test control: application to coagulation factor VIII assay.

Coagulation factor VIII was assayed around the critical concentration of 80 U/dl, which is optimal for postoperative haemostasis in haemophiliac patients, in order to assess the use of Bayesian logic in interpreting internal quality control results during a change of reagent or control batch. A mathematical model based on Bayesian inference, requiring no preliminary control-plan phase, was compared with a classical approach, which necessarily involves performing a preliminary phase. Tsiamyrtzis and Hawkins' Bayesian model proved applicable to rapid statistical control of factor VIII assay, detecting shift at least as efficiently as classical approaches, which depend on running the kind of costly and controversial preliminary control phase recommended by Shewhart.

Determining the adequate number of internal quality control levels: the example of coagulation factor VIII assay.

Taking the specific case of coagulation factor VIII assay, we determined the characteristics of an internal quality control panel assuring control of the assay method for all of the critical factor VIII concentrations. The precision of the assay method was determined on six control materials C1-C6, with expected factor VIII levels of 1, 5, 30, 50, 80 and 150 U/dl, respectively. Given that, when two control levels correlate statistically, the information provided by one of them is redundant, we used correlation and principal components analysis to define a priori adequate and inadequate control panels. For each of these panels, we calculated the number of runs required, using Hotelling's method, to detect a shift expressed on C1 and impacted on C2, C3, C4, C5 and C6 in relation to the correlation phenomena among the six levels. The C1/C6 panel proved to be as informative in this regard as the complete panel for a 1 U/dl shift simulated on C1 and impacted on other levels too. These correlation phenomena allow the biologist to implement fewer control levels than there are critical concentrations needing to be explored in the internal quality control plan.